[Intel-gfx] [PATCH 1/2] drm/i915: clear up backlight inversion confusion on gen4

Carsten Emde C.Emde at osadl.org
Mon Apr 23 18:55:37 CEST 2012


On 04/23/2012 05:56 PM, Daniel Vetter wrote:
> On Mon, Apr 23, 2012 at 05:38:27PM +0200, Carsten Emde wrote:
>> On 04/23/2012 05:22 PM, Daniel Vetter wrote:
>>> On Mon, Apr 23, 2012 at 05:06:53PM +0200, Carsten Emde wrote:
>>>> On 04/23/2012 04:22 PM, Daniel Vetter wrote:
>>>>> On Mon, Apr 23, 2012 at 04:00:23PM +0200, Carsten Emde wrote:
>>>>> [..]
>>>>> The idea was to boot with kms and see whether any of these values would
>>>>> restore the backlight. Writing to CTL1 should change anything in CTL2.
>>>> Ah, sorry, ok. Removed the quirk again and tested the various settings:
>>>>
>>>> ->   Initial screen: dark
>>>>
>>>> # intel_reg_write 0x61250 0x80000000
>>>> Value before: 0xE0000000
>>>> Value after: 0x80000000
>>>> ->   Still dark
>>>>
>>>> # intel_reg_write 0x61250 0xa0000000
>>>> Value before: 0x80000000
>>>> Value after: 0xA0000000
>>>> ->   Still dark
>>>>
>>>> # intel_reg_write 0x61250 0x90000000
>>>> Value before: 0xA0000000
>>>> Value after: 0x90000000
>>>> ->   BACKLIGHT ON!
>>>>
>>>> # intel_reg_write 0x61250 0xb0000000
>>>> Value before: 0x90000000
>>>> Value after: 0xB0000000
>>>> ->   Still ON.
>>>
>>> Neat. Let's test two more:
>>>
>>> # intel_reg_write 0x61250 0xd0000000
>>> # intel_reg_write 0x61250 0xc0000000
>> Here we go.
>>
>> ->   Initial screen: dark
>>
>> # intel_reg_write 0x61250 0xd0000000
>> Value before: 0xE0000000
>> Value after: 0xD0000000
>> ->   BACKLIGHT ON!
>>
>> # intel_reg_write 0x61250 0xc0000000
>> Value before: 0xD0000000
>> Value after: 0xC0000000
>> ->   Dark again.
>
> Ok, so the polarity bit does work as advertised. But I still don't
> understand how your machine works, so assuming your machine has backlight
> control keys,
Yes it has, but they don't have any effect.

Unfortunately, I have to leave now and will be away from the lab for two 
days. Will continue to work on the problem when I'll be back. For the 
time being, I let the two conditions, your new patch and the old quirk, 
coexist in the code.

Thanks,
	-Carsten.



More information about the Intel-gfx mailing list